Targets

If we're going to have large, centralised governments, we should probably accept that much of their policy will centre round the use of highly aggregated numerical data. The Social Policy Bond principle is no different, but I advocate that the bonds target broad indicators of wellbeing that are meaningful to ordinary people. Here's what happens when governments use Mickey Mouse micro-targets:

[H]ead teachers in the [UK's] government-controlled sector live and die by targets. Their careers depend on them. One of the targets is high attendance. Then came the snow [in the first week of February]. Most schools in London were closed on the Monday. But more schools managed to open on the Tuesday. Now you might think that those schools whose head teacher and staff struggled into work for the benefit of the children in their care were particularly conscientious. Not in target-land however. They were mugs. Since their schools were open, the many absences of children from their schools on that snowy day will count against them when it comes to calculating how good their attendance figures have been. But those schools which stayed shut will be deemed to have been shut because of exceptional circumstances and their attendance figures will not be affected at all. So the conscientious schools that opened will be punished with bad attendance figures. The less conscientious schools that stayed shut will be rewarded with high attendance figures. Thus does the state give perverse incentives. Source